Question Alder Lake - Official Thread

[Hulk]

With the release of Alder Lake less than a week away and the "Lakes" thread having turned into a nightmare to navigate I thought it might be a good time to start a discussion thread solely for Alder Lake.

 

[Heartbreaker]

@TESKATLIPOKA I believe Techspot is the website version of Hardware Unboxed.

Yes, this is the case. Same folks. Steve Walton is the author of the Article and he is the guy in the video.

 

[jpiniero]

https://hk.xfastest.com/135191/intel-core-i3-12100-cpu-test/

Alleged review of the 12100. Only compares to the 3300X and the 3100.

 

[TESKATLIPOKA]

@IntelUser2000 As you mentioned.
i7-6700K and i7-7700K are both 4C8T based on Skylake and Kabylake respectively, so they should have similar IPC to Gracemont, yet the difference in performance is 45% at 1% low or 25% at average FPS.

I also checked Techpowerup review, and they also tested Battlefield V, Cyberpunk 2077 and Shadow of the Tomb Raider.
8 P-cores no HT vs 8 E-cores, both at 3.9GHz. The difference between reviews was:

	Techspot	Techpowerup
Game	4x E-core vs 4x P-core	8x E-core vs 8x P-core no HT
Battlefield V	114 vs 213 (+87%)	169 vs 245 (+45%)
Cyberpunk 2077	71 vs 121 (+70%)	104 vs 116 (+12%)
Shadow of the Tomb Raider	73 vs 133 (+82%)	162 vs 251 (+55%)

The difference is much smaller in TPU review.

At TPU they also had i7-6700K, and It performed similarly to the E-cores, which is quite surprising considering It was 4C8T vs 8C8T and In Shadow of the Tomb Raider the i7-6700K was actually a lot faster 162 vs 236 FPS.

 

[coercitiv]

In my opinion, only 4 threads were a bottleneck. They could have also tested 8 E-cores.

Hyperthreading is probably the reason why the 6700K is much faster even though it's much closer(even identical) in other applications.

Remember the 6700K has 8MB L3 cache, while the 12600K has 20MB and 12900K has 30MB.

 

[LightningZ71]

The E-cores also share a single pool of L2, which negatively impacts their performance inhighly threaded situations.

 

[Hulk]

They are even on IPC

View attachment 52350

Anybody know where this chart came from?
Post #334

 

[igor_kavinski]

Anybody know where this chart came from?
Post #334

Different story in Cinebench 20: https://tweakers.net/reviews/9472/9...9-12900k-i7-12700k-en-i5-12600k-ipc-test.html

 

[Markfw]

Different story in Cinebench 20: https://tweakers.net/reviews/9472/9...9-12900k-i7-12700k-en-i5-12600k-ipc-test.html

View attachment 53697

No 5950x in that chart.

 

[Hulk]

No 5950x in that chart.

Doesn't matter, ST test with all CPU's at 3.5GHz.

 

[coercitiv]

Differences between Win11 and Win10, the 11 scheduler has subpar gaming performance with custom P/E core mixes, suggesting that the Win11 "optimization" for gaming is to always avoid the E-cores.



It's a beautifull mess, luckily we can drown any worry in gallons of Cinebench charts where scaling has zero issues.

 

[Shivansps]

So, keep using Windows 10?

 

[JoeRambo]

So, keep using Windows 10?

Disabling E cores works wonders too.

 

[Markfw]

Disabling E cores works wonders too.

So use Windows 10, because windows 11 is slower, or disable the ecores, which windows 11 was written for. These CPUs have serious issues IMO.

 

[xinthius]

So use Windows 10, because windows 11 is slower, or disable the ecores, which windows 11 was written for. These CPUs have serious issues IMO.

Hopefully there will be updates to Windows 11 scheduler in due course! Makes me think perhaps I should have gone for the 5600X rather than my 12600K.

That being said, I have excellent performance and no complaints thus far.

 

[coercitiv]

So, keep using Windows 10?

This testing shows Win11 cannot handle custom core configurations (yet), with stock 8+4 or 8+8 config performance will be fine in gaming. Let's wait for more testing from Hardawre Unboxed though, maybe we get some info from Intel too.

Disabling E cores works wonders too.

That's a good measure until the software side gets sorted out, but in the end customers would like to use all the cores they paid for.

 

[igor_kavinski]

I wonder how many 12900K owners are planning to drop the 13900K into their mobos when that CPU launches.

 

[coercitiv]

I wonder how many 12900K owners are planning to drop the 13900K into their mobos when that CPU launches.

For more E-core goodness? All of them, since they want the fatter L3 cache . If 13900K is 8+16 then you get 40MB L3 minimum, unless Intel goes further and pumps up the cache even more. "Improved cache for gaming", remember?

 

[JoeRambo]

That's a good measure until the software side gets sorted out, but in the end customers would like to use all the cores they paid for.

Not optimal ofc, but as consolation bonus you get extra L3 cache slices and none of the problems with scheduler. Also it improves performance of P cores by keeping uncore frequency higher, cause with E-Cores active, L3 cache can't be clocked that high, since uncore slices share voltage with E-Core L2 cluster.

So use Windows 10, because windows 11 is slower, or disable the ecores, which windows 11 was written for. These CPUs have serious issues IMO.

I might be a rarity - but i don't have any problems with Windows 11. Performance is great, task bar ( after moving it to the left side ) is fine. But OS definitely needs more work, there are some features that are lacking atm, but might get fixed in the future.

 

[tomatosummit]

These performance problems are a good indication of why intel went with 6p cores in the mainstream desktop segment.
Specifically the low end i5s marketed for budget gaming, it's a segment that just needs to plug in, add a graphics card and work well enough among a wide range of systems and software with users who will not be able to troubleshoot these things.

While I still anticipate the 2p+8e core cpus I'm not sold on desktop e-cores. They feel like marketing gimmick today.
Give me 2-4 for low power draw in idle and tertiary tasks or a huge 32+ array for specialised workstation tasks.
I think it's telling how apple's m1++ is laid out. 8cores for performance and 2 efficient for extended battery life that's letting them scale across performance laptops, DTR and mid range workstations (big imacs).

 

[Heartbreaker]

Differences between Win11 and Win10, the 11 scheduler has subpar gaming performance with custom P/E core mixes, suggesting that the Win11 "optimization" for gaming is to always avoid the E-cores.

It's a beautifull mess, luckily we can drown any worry in gallons of Cinebench charts where scaling has zero issues.

So, keep using Windows 10?

In that chart it looks like there's only a problem here, if you create one by disabling P cores. It's very difficult to find an actual problem if you just leave the CPU with all it's cores enabled (6 to 8 P cores on all shipping CPUs). What video is the source of that chart??

What Intel is doing works fine, and actually makes some sense. Since they aren't shipping any low P core parts aimed at gaming. Only the ultra low power mobile part has less than 6 P cores.

With 6 P cores devoted to gaming, the E cores will deal the background tasks.

 

[Hulk]

These CPUs have serious issues IMO.

You recently chastised someone for making a hyperbolic statement regarding 12900K performance vs. 5950X. In all fairness, do you think your statement is a little over the top? And to quote you, "do you own one?"

Games need certain threads to stay on the P cores and the scheduler in rare cases isn't doing this correctly. It's a pretty simple press of the scroll lock to disable the E's in those rare cases.

I have a 12700K and don't game. I have ZERO issues. In fact, this is the most stable system I've ever built.

Besides the game issue I mentioned, which I don't think qualifies as "serious issues," what other issues are you referring to?

 

[TheELF]

Differences between Win11 and Win10, the 11 scheduler has subpar gaming performance with custom P/E core mixes, suggesting that the Win11 "optimization" for gaming is to always avoid the E-cores.

View attachment 53700

It's a beautifull mess, luckily we can drown any worry in gallons of Cinebench charts where scaling has zero issues.

You can't have blanket conclusions based on one single game.
Tomb raider was also the poster game to show how bad it is to have two ccxs, it took them a while to fix that.

 

[TheELF]

Games need certain threads to stay on the P cores and the scheduler in rare cases isn't doing this correctly. It's a pretty simple press of the scroll lock to disable the E's in those rare cases.

The scheduler is working fine putting low priority threads or threads they deem low priority by the instructions they use on e cores, and rendering graphics which is what a game would be doing checks both these requirements.
Games have to be updated to tell the OS that game rendering is pretty important, no scheduler no matter how smart can tell the difference between a game rendering graphics and cinebench or whatever other background task rendering graphics.

 

[jpiniero]

What Intel is doing works fine, and actually makes some sense. Since they aren't shipping any low P core parts aimed at gaming. Only the ultra low power mobile part has less than 6 P cores.

I'm sure there will be plenty of 4+8 laptops that will at least come with a dGPU.

 

[Heartbreaker]

I'm sure there will be plenty of 4+8 laptops that will at least come with a dGPU.

There are no 4+8 Alder Lake CPUs at all. None even rumored.